Shadowing means speaking along with a recording, a fraction of a second behind it, like a shadow. It is simple, but it trains something ordinary "listen and repeat" cannot: the timing, reductions and melody of connected speech.
The reason is memory. When you hear a whole sentence and then repeat it, your memory keeps the words and quietly throws away the rhythm, so you reproduce the sentence in your own accent. Shadowing leaves no gap for that to happen. This step-by-step guide covers what shadowing is and is not, a four-stage routine for a single clip, how to choose material, and what to do when it is not working. For a longer treatment with a 30-day plan, see the complete shadowing guide.
Shadowing is not repetition
Shadowing means speaking along with an audio recording almost simultaneously, trailing the speaker by a fraction of a second, without waiting for pauses. This is what distinguishes it from the listen-and-repeat exercise familiar from classrooms, and the difference matters.
When you repeat after a pause, you reproduce your memory of what you heard, and memory normalises: it stores the words and discards the timing, the reductions and the pitch. When you shadow, there is no time for memory to intervene. You are copying the acoustic signal as it arrives, including everything your memory would otherwise have thrown away.
| Listen and repeat | Shadowing | |
|---|---|---|
| Timing | After a pause | Simultaneous, trailing by a fraction of a second |
| What you copy | Your memory of the words | The signal itself |
| Rhythm | Reverts to your own | Constrained by the model |
| Reductions | Usually lost | Preserved, because there is no time to expand them |
| Main benefit | Individual sounds | Rhythm, linking, intonation |
Four stages with one clip
Use a single clip of 20 to 30 seconds and work through all four stages on it, rather than shadowing a new clip each day. Repetition on the same material is where the gains are.
- Silent shadowing. Listen three or four times while mouthing the words without sound. This builds the articulatory pattern without the distraction of hearing yourself.
- Mumbling. Shadow at very low volume, barely voiced. The goal here is timing alone — staying in sync — not accuracy.
- Full shadowing with the script. Speak at normal volume while reading the transcript. Mark the linking points and the stressed syllables on the script first.
- Full shadowing without the script. The real test. If you can stay in time without reading, the pattern has been internalised rather than decoded.
Choosing material
Material selection determines whether shadowing works, and most failures trace back to a clip that was wrong for the stage.
- Length: 20 to 30 seconds. Longer clips cannot be repeated enough times in a session.
- Type: unscripted speech — interviews, podcasts, documentary narration. Scripted news reading is unnaturally even, and graded classroom audio removes exactly the reductions you want to learn.
- Difficulty: you should understand it comfortably when listening normally. Shadowing is not a comprehension exercise, and using material you do not understand turns it into decoding.
- Accent: one accent, consistently. Alternating between American and British models produces a mixture that neither group recognises.
- Speaker: if pitch matters to you, choose a speaker whose vocal range is similar to yours, so that copying the intonation does not require an uncomfortable register.
Ten repetitions of one clip beat one repetition of ten clips. The benefit comes from the pattern becoming automatic, which requires repetition of the same material. Learners who change clips daily get variety and very little improvement.
What shadowing is good at, and what it is not
Shadowing is the most efficient exercise available for rhythm, linking, weak forms and intonation, because all four are properties of timing and the model enforces the timing for you.
It is much weaker for individual sounds. If you substitute /s/ for /θ/, shadowing will not correct it — you will simply produce your substitution in the right rhythm. Segmental errors need targeted work with explicit articulatory instructions and feedback.
The sensible combination is targeted drills for the two or three sounds you are currently correcting, plus shadowing for everything that drills cannot reach.
Common problems
| Problem | Likely cause | Fix |
|---|---|---|
| You fall behind constantly | Clip too fast, or you are reading ahead | Slow playback to 0.8x; return to mumbling stage |
| You can only do it with the script | You are decoding, not shadowing | More repetitions at stage two before removing the script |
| It feels mechanical and boring | Same clip for too long after mastery | Change the clip once you can do stage four twice |
| Your accent has not changed | Shadowing only, no segmental work | Add targeted drills; shadowing does not fix individual sounds |
| You get tired quickly | Normal at first — it is physically demanding | Start at five minutes and build up |
Frequently asked questions
How is shadowing different from listening and repeating?
In shadowing you speak along with the recording almost simultaneously, trailing by a fraction of a second, without waiting for pauses. Repeating after a pause reproduces your memory of what you heard, and memory normalises — it keeps the words but discards the timing, reductions and pitch. Shadowing gives memory no time to intervene, so you copy the acoustic signal including everything memory would have thrown away.
How long should a shadowing clip be?
Twenty to thirty seconds, used repeatedly rather than replaced daily. Ten repetitions of one clip produce far more benefit than one repetition each of ten clips, because the gain comes from the pattern becoming automatic. Learners who change material every day get variety and very little measurable improvement.
What kind of audio works best for shadowing?
Unscripted natural speech such as interviews, podcasts or documentary narration, at a level you already understand comfortably. Scripted news reading is unnaturally even, and graded classroom audio removes exactly the reductions and linking you are trying to learn. Stick to one accent, since alternating between American and British models produces a mixture neither group recognises.
Will shadowing fix my individual sound errors?
Generally no. Shadowing is the most efficient exercise available for rhythm, linking, weak forms and intonation, because the model enforces the timing. But if you substitute /s/ for /θ/, shadowing will simply place that substitution in the correct rhythm. Individual sounds need targeted drills with explicit articulatory instructions, so the two approaches are complementary rather than interchangeable.
Disclosure: this page links to Preply, a tutoring marketplace. If you book a lesson through that link we may earn a commission, at no extra cost to you. It does not change what is recommended here. See our disclaimer and privacy policy.
Shadow one clip today
Pick a clip, run the four stages from this guide, and record the last pass.
Open shadowing practice